Grok AI: MCMC Wants To Sue X Over Failure To Ensure User Safety

Subscribe to our Telegram channel for the latest stories and updates.

The Malaysian Communications and Multimedia Commission (MCMC) is seeking to haul X Corp. (formerly Twitter) and xAI LLC to court for failing to ensure user safety in relation to their artificial intelligence (AI) tool, Grok.

According to a statement MCMC published yesterday (13 January), it said that solicitors had been appointed and legal proceedings would commence soon.

“MCMC has identified the misuse of Grok to generate and disseminate harmful content, including obscene, sexually explicit, indecent, grossly offensive and non-consensual manipulated images.

“Content allegedly involving women and minors is a serious concern. Such conduct contravenes Malaysian law and undermines the entities’ stated safety commitments,” said the commission.

The statement also said that X Corp. and xAI LLC are responsible even though the content was generated by users.

“They retain control of the design, implementation, supervisory mechanism, as well as risk mitigation steps for Grok. Responsibility cannot be ignored when safety fails,” it added.

Notices were given to X Corp. and xAI LLC but no action was taken

It was earlier reported that notices were issued to X Corp. and xAI LLC on 3 and 8 January to remove offensive content, but no remedial actions were taken.

On 11 January, MCMC placed a temporary block on Grok for all Malaysian users following the Indonesian government’s block on the AI for its citizens.

The issue stems from users who use the Grok AI to generate sexual deepfake images without consent.

Deepfake is realistic, AI-generated synthetic media (video, audio, or image) that convincingly depicts a person saying or doing something they never did, created using deep learning to swap faces, manipulate expressions, or synthesize voices.

At time of press, the Grok website is inaccessible but the AI can still be used via X on mobile or desktop browser.

Are there laws against generating deepfake images in Malaysia?

Section 233 of Communications and Multimedia Act 1998 criminalizes “improper use of network facilities” to create or transmit content that is obscene, indecent or offensive. The punishment under the Act is a fine of up to RM50,000, imprisonment of up to one year, or both.

If the images are used to harass, insult, or extort a victim, the Penal Code provides harsher penalties:

Misuse of deepfake is also governed under the Sexual Offences Against Children Act 2017, whereby if the deepfake involves a minor (anyone under 18), the penalties are severe, as AI-generated Child Sexual Abuse Material (CSAM) is treated the same as real images. The punishment for creating such material includes up to 30 years in prison and at least six strokes of the rotan (whipping).

...